Concentrative Nucleoside Transporter (rCNT1) Is Targeted to the Apical Membrane through the Hepatic Transcytotic Pathway

Experimental Cell Research 281, 77– 85 (2002) doi:10.1006/excr.2002.5641

Concentrative Nucleoside Transporter (rCNT1) Is Targeted to the Apical Membrane through the Hepatic Transcytotic Pathway Sylvie Duflot,* Maria Calvo,† F. Javier Casado,* Carlos Enrich,† and Marc¸al Pastor-Anglada* ,1 *Departament de Bioquimica i Biologia Molecular, Facultat de Biologia, and †Department de Biologia Celular, Institut d’Investigacions Biome`diques August Pi i Sunyer (IDIBAPS), Facultat de Medicina, Universitat de Barcelona, E-08028 Barcelona, Spain

The hepatocyte is a highly polarized epithelial cell with a plasma membrane divided into three major functional domains: sinusoidal, facing the blood and the hepatic endothelial cells; the lateral domain, containing the junction complexes (e.g., desmosomes and gap junctions), and the canalicular plasma membrane, involved in bile secretion [8]. The origin and maintenance of hepatocyte functional polarity depends essentially on the membrane trafficking events, exocytosis, endocytosis, transcytosis, and recycling, which in part are governed by two main distribution centers involved in the sorting of molecules: the TGN and the early/ sorting endosome (CURL). The targeting of membrane proteins to the hepatic canalicular plasma membrane was classically believed to be undertaken by the indirect route (transcytosis), via basolateral delivery and the subsequent transport through the endocytic structures [9, 10]. However, a direct route from the TGN to the canalicular membranes was considered an alternative pathway, and this has recently been demonstrated for canalicular ABC transporters [11]. Nucleoside transporter activity in plasma membrane vesicles partially purified from the sinusoidal domain did not show any particular enrichment of the CNT1 protein, although it appears to bear significant amounts of the CNT2 isoform [5]. The possibility that nucleoside transporters are also located at the canalicular plasma membrane was suggested by others on the basis of a special need for salvage of the nucleosides generated by canalicular ecto-ATPases [12] and has been demonstrated for CNT1 in rat liver [13]. Recently, it has been demonstrated that, in the sorting of membrane proteins at the TGN, hepatocytes differ from other epithelial cells (e.g., MDCK cells) since they do not express an epithelial cell specific subunit (␮1B) of the AP1 clathrin adaptor complex [14, 15]. The lack of ␮1B correlates with the indirect pathway, which is the major sorting pathway in hepatocytes. Here we examine the targeting of CNT1, first, by monitoring nucleoside transport properties and CNT expression in isolated basolateral and canalicular membranes; second, by determining CNT protein

The Na ⴙ-dependent nucleoside transporter CNT1 has been identified in a caveolin-enriched plasma membrane fraction (CEF), in transcytotic endosomes, and in canalicular membranes isolated from quiescent rat liver in which the transporter appears to be biologically active. CNT1 was also detected, albeit in small amounts, in the early/sorting endosomes. Plasma membrane preparations enriched in basolateral markers showed Na ⴙ-dependent nucleoside transport activity that is mostly, if not exclusively, accounted for by CNT2, a transporter protein which was not detected in CEF nor in the endosomal fractions. These data are consistent with different localization and trafficking pathways of the two isoforms in hepatocytes. CNT1 is the first transporter which is reported to follow the transcytotic pathway to be inserted on the apical side of liver parenchymal cells. © 2002 Elsevier Science (USA)

Key Words: concentrative nucleoside transporter (CNT); hepatocyte; endosome; endocytosis; transcytosis; caveolin.

INTRODUCTION

Natural nucleosides and nucleoside-derived drugs used in anticancer and antiviral therapies are transported into cells by a variety of plasma membrane carriers belonging to two gene families, CNT and ENT [1– 4]. CNT1- and CNT2-related proteins are responsible for the concentrative Na ⫹-dependent high-affinity transport of pyrimidine and purine nucleosides, respectively [1– 4]. These carrier proteins may be coexpressed in a single cell type and may undergo differential regulation, as reported in rat hepatocytes [5, 6]. A third member of this family showing broad substrate specificity, CNT3, has been recently cloned from human and mouse tissues [7]. 1

To whom correspondence and reprint requests should be addressed at Departament de Bioquimica i Biologia Molecular, Universitat de Barcelona, Diagonal 645, E-08028 Barcelona, Spain. Fax: 34-93-4021559. E-mail: [email protected]. 77

0014-4827/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

78

DUFLOT ET AL.

amounts and distribution, using polyclonal monospecific antibodies, in highly purified endosomes and caveolae-enriched plasma membrane fraction from rat liver; and, third, by immunogold electron microscopy. MATERIALS AND METHODS Isolation of endosomes, plasma membrane, and caveolae-enriched fractions from rat liver. Basolateral plasma membranes initially used to identify CNT1 and CNT2 proteins and their related transport activities were purified from adult rat livers by a Percoll-density method as described [16]. Canalicular plasma membrane fractions were isolated separately [17]. The two types of membrane fraction were then washed twice in a Hepes buffer containing isotonic sucrose, frozen in liquid nitrogen, and stored at ⫺80°C. Plasma membrane enrichments versus homogenates were calculated from 5⬘nucleotidase enzyme activity measurements, as previously described [16]. The isolation of the endosomal fractions involved previous treatment of the rats for 3 days with 17-␣-ethinyl estradiol, which induces the expression of the low-density lipoprotein (LDL) receptors. Then, rats were anesthetized with isofluorane, and human LDL (5 mg of protein) was injected into the femoral vein. Twenty minutes later livers were removed and homogenized in 0.25 M sucrose in the presence of protease inhibitors. Three distinct endosomal fractions were obtained after centrifugation of a crude endosome fraction in a sucrose gradient: MVB at 8.24%/19.3%, CURL at 19.3%/28.81%, and RRC at 28.81%/36.37% (w/v) interfaces [18, 19]. In this procedure, a fourth band at the interface of 36.37%/46%, just on the cushion of heavy sucrose, was also collected and characterized. Due to its considerable enrichment in caveolin-1 and its morphology, this fraction was called CEF (from caveolin-enriched plasma membrane fraction). This fraction derives from the sinusoidal plasma membrane of the hepatocyte and has been characterized biochemically and morphologically [19, 20]. Each fraction was collected and ice-cold water was added to render the fractions isotonic. These fractions were then pelleted, resuspended in 0.9% NaCl, and stored at ⫺80°C. Nucleoside transport activity in basaloteral and canalicular plasma membrane vesicles. Nucleoside uptake was measured as previously described, using a rapid filtration technique [21]. One micromolar uridine, guanosine, or cytidine transport rates were assessed at very short incubation times (3 s), under, initial velocity conditions, by mixing 20 ␮l of membrane vesicles (10 –20 ␮g protein/ ␮l) with 20 ␮l of the incubation mixture in Eppendorf tubes to give the following concentrations: 0.25 M sucrose, 0.2 mM CaCl 2, 10 mM MgCl 2, 10 mM Hepes/KOH, pH 7.4, 100 mM either Na or K-sulfocyanate (SCN), and the tritiated substrate (routinely 1 ␮Ci per incubation). The whole content of the tube was then passed through a nitrocellulose filter (Millipore, pore size 0.45 ␮m) and washed several times with a cold stop solution. Filters were counted for radioactivity. Western blot analysis of the membrane fractions. Monospecific polyclonal antibodies against the rat CNT1 and CNT2 isoforms have been characterized in our laboratory [5, 22]. The following antibodies were commercially available: anticaveolin 1 (Transduction Laboratories) and Rab 5 (Santa Cruz). The antibody against ASGP-R was raised in our laboratory [23]. Antibodies against the bile acid transporters Ntcp and Bsep had been previously characterized [24] and were a kind gift from Bruno Stieger (University Hospital, Zurich, Switzerland). The antibodies against 5⬘-nucleotidase, pIgR, and cellubrevin were kind gifts from Dr. Paul Luzio (Cambridge, UK), Dr. K. E. Mostov (San Fransisco, CA), and Joan Blasi (Barcelona, Spain), respectively, and have been previously used in our laboratory [25]. Five to 10 ␮g of protein was routinely run on a 10% polyacrylamide– SDS gel. Proteins were transferred onto filters (Immobilon-P, Millipore, 0.45-␮m pore size), which were then blocked overnight in a 5%

dry-milk-supplemented, 0.2% Tween 20, phosphate-buffered saline solution prior to immunoreaction. Filters were then incubated for 1 h with the diluted antisera. Dilutions used were as follows: 1/2000 for CNT1, 1/1000 for CNT2, 1/5000 for Ntcp, 1/2000 for Bsep, 1/2000 for caveolin, 1/400 for ASGRP and pIgR, 1/100 for Rab5, 1/200 for cellubrevin, and 20 ␮g/ml for 5⬘-nucleotidase. A goat anti-rabbit IgG coupled to horseradish peroxidase was used as a second antibody and diluted 1/2000 in dry-milk-supplemented 0.2% Tween 20 PBS. ECL (Amersham) was used to visualize the proteins. Immunogold electron microscopy. Livers of Wistar rats were perfused in situ with PBS to wash off blood and then fixed with PBS containing 4% paraformaldehyde, pH 7.4. Liver pieces were placed in cold fixative and further dissected. The tissue was first washed in PBS containing 0.02 M glycine at 4°C for 2 h and then cryoprotected, also in the cold, with subsequent incubations in PBS with increasing concentrations of sucrose (0.6 M for 2 h, 1.3 M for 4 h, and, finally, 2.3 M overnight). Ultrathin sections were cut using a Leica Ultracut UCT, collected onto Formvar-coated gold, and stored on 2% gelatin/ PBS at 4°C. For CNT1 and CNT2 localization, grids were sequentially incubated at room temperature for 10 min in PBS, containing 2% gelatine, pH 7.4, for 3 min in PBS containing 0.02 M glycine, and finally for 10 min in the same solution supplemented with 0.1% FCS. Primary antibodies were used at a dilution of 1/100 and 1/50 for rCNT1 and rCNT2, respectively, in 0.02 M glycine, 0.1% FCS PBS. Grids were incubated for 1 h and then washed in the same buffer before a second incubation was begun, using Protein A conjugated to colloidal gold particles, diluted in the same solution, and kept for 45 min at room temperature. After several washes, sections were stained with 2% methylcellulose, 0.2% uranylacetate, and examined using a transmission electron microscope JEOL 1010. As control for single immunostaining, sections were incubated with Protein A– gold only. The labeling was specific, as no signal was obtained in control. Immunofluorescence. Livers were perfused and small pieces were cryoprotected with subsequent incubations in PBS with increasing concentrations of sucrose (10% for 2 h, 20% for 4 h, and, finally, 30% overnight, cryostat sections (6 – 8 ␮m) were obtained, air-dried, and hydrated in PBS before immunostaining. The antiserum against rCNT1 was used at a dilution of 1/200, whereas the antibody against caveolin 1 (IgM) was used at 1/10. Texas-red-labeled anti-mouse and Oregon-green-anti-rabbit antibodies were used as secondary antibodies at 1/50. Prebleed serum for rCNT1 and secondary antibodies alone were used as negative controls.

RESULTS

CNT1 and CNT2 Proteins in Canalicular and Basolateral Plasma Membrane Fractions Plasma membrane fractions enriched in canalicular and basolateral domains of the hepatocyte were purified and the amount of CNT1 and CNT2 proteins was monitored by Western blot analysis. Nucleoside transport activities associated with these two membrane fractions were also examined. The purity of these preparations was initially determined by measuring their relative enrichments in 5⬘-nucleotidase enzyme activity, as well as in well-characterized markers of canalicular and basolateral domains: the bile acid transporters Bsep and Ntcp, respectively, as detected by Western blot analysis. 5⬘-Nucleotidase activity mean enrichments were 14- and 28-fold for canalicular and basolateral membranes, respectively. Figure 1 shows representative Western blot analyses of crude homog-

CNT1 FOLLOWS THE HEPATIC TRANSCYTOTIC PATHWAY

FIG. 1. rCNT1 and rCNT2 in basolateral and canalicular membranes from rat liver. Representative Western blots of rat liver homogenates (Hb and Hc), basolateral membrane (Vb), and canalicular membrane (Vc) preparations, using antibodies against either rCNT1 or rCNT2, are shown. Hb and Hc refer to crude tissue homogenates obtained for basolateral and canalicular plasma membrane preparations, respectively. Antibodies against Bsep and Ntcp were used as canalicular and basolateral markers, respectively.

enates and basolateral and canalicular membrane fractions, using antibodies against CNT1, CNT2, Bsep, and Ntcp. Ntcp and Bsep were not detected in crude homogenates but were abundant in basolateral and canalicular fractions, respectively. This means that the two fractions were enriched not only in plasma membrane, as deduced from the 5⬘-nucleotidase data, but also in their respective functional domains. No Bsep was detected in basolateral membranes, nor was Ntcp

79

found in significant amounts in canalicular membranes, thus suggesting that cross contamination was very low or even negligible. CNT2 was barely detected in crude homogenates and canalicular membranes, but it was abundant in basolateral fractions. Due to its negligible presence in homogenates a reliable enrichment factor could not be calculated, although this has no bearing on the observation that CNT2 is highly expressed in basolateral plasma membranes of hepatocytes. CNT1 was detected in crude homogenates and present, although to a lesser extent, in plasma membrane fractions from the two origins, basolateral and canalicular. The average amount of CNT1 in vesicles was routinely less than one-third of that found in crude homogenates. This observation suggested, as will be demonstrated below, that a significant amount of CNT1 is located intracellularly in rat hepatocytes. In accordance with this pattern, Na ⫹-dependent uptake of uridine into basolateral vesicles appeared to be mediated mostly, if not exclusively, by CNT2. Uptake of guanosine (a CNT2 preferring substrate) was Na ⫹dependent (Fig. 2A), whereas neither transport of cytidine (a CNT1-preferring substrate) (Fig. 2A) nor inhibition of uridine uptake by cytidine (not shown) was detected. Plasma membrane vesicles showed relatively high Na ⫹-independent nucleoside uptake rates. This is a common feature of this in vitro model, even when using other substrates, such as amino acids [16]. In any case, as previously reported [21], this Na ⫹-independent nucleoside transport activity is mostly NBTIinsensitive. Canalicular membranes retained Na ⫹-dependent nucleoside transport activity which was consistent with the expression of, at least, a CNT1-type transporter, as deduced from the fact that, in contrast to the basolateral fractions, concentrative uptake of cytidine was detected at similar levels to that found for uridine (Fig. 2B). However, these vesicles also showed significant Na ⫹-dependent guanosine uptake (Fig. 2B). Interestingly, this transport activity could not be fully inhibited by high concentrations of cytidine (less than 30% inhibition at 100 ␮M cytidine) (not shown). This purine-preferring transport activity is unlikely to be associated with CNT2, since low amounts of CNT2 protein were found in canalicular vesicles and uptake would not be inhibited by a pyrimidine nucleoside. Thus, it may be the result of the expression of the recently cloned new isoform CNT3, which appears to be expressed in the liver, although it has not yet been found in parenchymal cells. Detection of the CNT1 Transporter in Intracellular Membrane Fractions In the hepatocyte, the apical targeting of membrane proteins is still unclear and two different pathways are

80

DUFLOT ET AL.

FIG. 2. Nucleoside transport activities in basolateral and canalicular plasma membranes from rat liver. Basolateral (A) and canalicular (B) membrane vesicles isolated from rat liver were tested for Na ⫹-dependent nucleoside transport activity, using tritiated uridine, guanosine, and cytidine as substrates (all at 1 ␮M). Uptake was measured, as indicated under Materials and, Methods, either in a NaSCN or in a KSCN medium. Na ⫹-dependent nucleoside transport was then calculated by subtracting the uptake rates in choline buffer from those measured in sodium medium. Guanosine is a CNT2specific substrate, whereas cytidine is a CNT1-specific substrate. Results are the mean ⫾ SD of five measurements made on pooled vesicles from four rats.

operational: the direct pathway from the TGN to the canalicular plasma membrane, only demonstrated for the ABC transporters, and the indirect pathway, which involves the endocytic/transcytotic compartment. To find out whether the indirect route is involved for the CNT1 transporter, we isolated endocytic fractions from rat liver homogenate. Three endosomal fractions— MVB (multivesicular bodies), CURL (compartment of uncoupling receptors and ligands— early/sorting endosomes), and RRC (receptor-recycling compartment)— and a caveolae-enriched plasma membrane fraction, recently characterized and derived from the sinusoidal plasma membrane domain of hepatocytes, were isolated. These fractions have been comprehensively studied by biochemical and morphological means [23, 25]. Figure 3A shows representative Western blots of markers used to define these fractions. Cellubrevin and Rab5, as early endocytic markers, polymeric immunoglobulin receptor (pIgR) and 5⬘-nucleotidase as transcytotic proteins, and the ASGP-R as a recycling receptor were also present in the MVB and CURL. Caveolin is a marker protein for CEF [25]. Caveolin was considered a marker of structures enriched in CEF. Electron microscopy and biochemical studies showed that these structures derived from the sinusoidal plasma membrane. In fact, CEF was the only fraction containing the scavenger receptor SR-BI [19], the physiological HDL receptor that mediates the selective uptake of cholesterol esthers; SR-BI is enriched in hepatocytes (as well as in adrenal, testis, and ovary) and localized in caveolae [26, 27]. Caveolin was detected, by Western blotting, in the early/sorting endosomes (CURL), RRC, and CEF but

FIG. 3. rCNT1 distribution in endosomal fractions and a caveolin-enriched plasma membrane fraction from rat liver. A: Three endosomal fractions, MVB, late endosomes, CURL, early endosomes, RRC, receptor-recycling compartment, and the caveolin-enriched plasma membrane fraction (CEF) were analyzed biochemically using a set of markers. A representative identification of selected proteins by Western blot analysis is shown. Cellubrevin and Rab5 are endosomal markers. plgR and ASGPR are hepatic receptors present in endosomes; 5⬘nucleotidase (5⬘NT), a GPI-anchored protein, is enriched in RRC. B: rCNT1 and rCNT2 were analyzed in all these fractions. A representative Western blot analysis shows the distribution of the rCNT1 isoform. rCNT2 was not detected in these subcellular fractions.

CNT1 FOLLOWS THE HEPATIC TRANSCYTOTIC PATHWAY

81

FIG. 4. Immunolocalization of rCNT1 in CEF, endosomal vesicles, and canalicular membranes. Immunoelectron micrographs of ultrathin sections of rat liver labeled with the monospecific polyclonal rabbit rCNT1 antibody, using cryoimmunoelectronmicroscopy, are shown. The basal domain of the membrane indicated in A is further amplified in B, showing the occurrence of rCNT1 in structures at the plasma membrane as well as in early endosomal vesicles (indicated by arrows). C shows the apical location of rCNT1, according to its functional activity in canalicular rat liver membrane preparations. No rCNT2 was found in the apical side of hepatocytes.

not in the late endocytic fraction (MVB). Although caveolin was more enriched in RRC than in CEF over the postnuclear supernatant (10.5 ⫾ 3 and 4.6 ⫾ 2, respectively), CEF (with 61%) was the fraction containing the highest amount of caveolin, followed by RRC (35%) and CURL (4%), when the total amount of protein recovered in each fraction was considered [25]. An identical fraction was isolated and described by Schnitzer et al. [28]. These endosomal fractions have been isolated and characterized in detail elsewhere [29 –32]. The presence of CNT1 and CNT2 proteins was studied in these endosomal fractions and in CEF by Western blotting. As shown in Fig. 3B, the pyrimidinepreferring nucleoside transporter CNT1 was found in the three endosomal fractions and CEF, although in different amounts. CNT2 was not detected in these membrane preparations with the available antibody. Mean enrichment values for CNT1 in these subcellular fractions versus crude homogenate were as follows: RRC, 6.3; CEF, 3.0; CURL, 0.8; MVB, 0.4. These data are consistent with the hypothesis that a significant amount of CNT1 protein is located intracellularly. The different subcellular localization of CNT1 and CNT2 was further analyzed by immunoelectron microscopy (Figs. 4 and 5). Figure 4A shows the basal domain of a hepatocyte (amplified in Fig. 4B), in which CNT1 was detected both in vesicular structures (indicated with an arrow), consistent with its presence in early endosomes, and in invaginations of the membrane, in agreement with its detection in CEF. Figure 4C shows the presence of CNT1 in the microvilli on the

apical side of a hepatocyte (canalicular plasma membrane). Figure 5 shows the basolateral membrane location of the CNT2 protein at the microvilli. The number of gold particles associated with membranes was counted on 50 fields from 10 sections. Data derived from these analysis are shown in Table 1. As previously pointed out, most of the CNT1 appears to be located intracellularly, but a significant amount is on the apical side. Most of the CNT2 is present on the basolateral side. Overall, these data are consistent with the results of subcellular fractionation as well as with the biological activity detected at the two poles of the hepatocyte. Finally, to further analyze the putative proximity of CNT1 to caveolae, double immunohistochemistry was performed on paraformaldehyde-fixed frozen sections of rat liver. As shown in Fig. 6 caveolin and CNT1 stains did not merge, which was as predicted because there is no functional basis to expect colocalization of these two proteins (also, CNT1 does not contain the motif for binding to the scaffolding domain of caveolin). Nevertheless, CNT1 was located close to caveolin and, therefore, we cannot rule out that CNT1 might be in raft microdomains in the close proximity of GPI-anchored proteins (destined to the apical membrane). DISCUSSION

Here we examine the evidence that hepatocytes express at least two isoforms of concentrative nucleoside transporters, CNT1 and CNT2 [5, 33], whereas basolateral plasma membranes are not equally enriched in

82

DUFLOT ET AL.

FIG. 5. Immunolocalization of rCNT2 in basal membranes. Immunoelectron micrographs of ultrathin cryosections of rat liver labeled with the monospecific polyclonal rabbit rCNT2 antibody. A shows the area further enlarged in B, corresponding to basal microvilli in which rCNT2 was identified (arrows).

these two transporters [5, 21], which suggests different subcellular locations for these proteins. Moreover, the possibility that CNT1 is targeted to the apical side of the hepatocyte, recently supported by evidence showing CNT1 staining in rat liver canaliculi [13], prompted

TABLE 1 rCNT1 and rCNT2 Distribution in Hepatocyte Membranes Plasma membrane CNT1 Basolateral Apical CNT2 Basolateral Apical

Au, % of total

Intracellular vesicles

Intracellular membranes

Other

6.25 35

37

18.8

3

84 N.D.

2

4

10

Note. rCNT1 and rCNT2 were detected in rat liver cryosections immunolabeled with anti-CNT1 and anti-CNT2 antibodies and Protein A gold (10 nm). Sections were examined, digitized, and photographed using a JEOL JEM 1010 electron microscope. A total of 310 (210,CNT1; 100, CNT2) gold particles associated with membranes were counted, corresponding to 50 fields from 10 sections. Au (% of total) refers to the percentage of the total gold particle amount found in a particular subcellular domain. Gold particles associated with the plasma membrane (sinusoidal, basolateral or canalicular, apical) was distinguished from labeling in intracellular vesicles or intracellular membranes or associated with nonparenchymal cells or in the extracellular matrix components (other). In a more detail examination it was evident that CNT2 almost exclusively labels microvilliar structures of the sinusoidal domain of hepatocytes (84%); however, CNT1 labels the apical membrane (35%) as well as a variety of intracellular vesicles (37%).

us to study the trafficking pathways involved in the insertion of CNT1 at the canalicular membrane. The study of the intracellular pathways involved in the trafficking of the CNT proteins in hepatocytes is important for the understanding of liver physiology, since the two isoforms appear to be up-regulated in the early phases of liver growth after partial hepatectomy [5, 34]. CNT2, at least, may be considered a cell cycle regulated gene [6]. The two transporter proteins are developmentally regulated and CNT2 expression is increased in primary cultures from fetal rat hepatocytes induced to differentiate in culture [35]. Interestingly, CNT2 expression is impaired in chemically induced hepatocarcinomas, whereas CNT1 is not detected in lesions from Alb-SV40 transgenic rat livers [22]. Moreover, both CNT1 and CNT2 are also coexpressed in other epithelia, such as enterocytes and renal proximal tubule cells. They are inserted at the apical side but neither cell type has been examined in an attempt to understand the trafficking pathways for these transporters in absorptive epithelia. This contribution is the first attempt to elucidate trafficking pathways of CNT transporters in polarized cell types. Our results indicate that CNT2 is mostly basolateral whereas CNT1 is essentially canalicular, although most of the latter appears to be located intracellularly, at least under steady-state conditions. The insertion of CNT transporters in these two domains of the plasma membrane of the hepatocyte is consistent with the biological activity of nucleoside transport found in these two fractions. In agreement with this subcellular

CNT1 FOLLOWS THE HEPATIC TRANSCYTOTIC PATHWAY

83

This enrichment is consistent with the proposed role of RRC in transporting specific transcytotic proteins, such as GPI-anchored proteins, to the apical surface. In addition, it seems clear that CNT1 takes the indirect route to the apical domain (no TGN–Golgi markers were detected in CEF) [25]. The putative trafficking of CNT1 in hepatocytes is summarized in Fig. 7. CNT1 and CNT2 transporters may be initially inserted into the sinusoidal plasma membrane, but at different microdomains, since CNT2 was not identified in the CEF fraction. This event may determine their functional status. CNT1 may be located in the close proximity of the caveolae and then be internalized into the early endosomes and eventually translocated into the transcytotic structures, prior to its insertion into the apical plasma membrane domain. CNT2 was present in negligible amounts in these fractions, although its activity appeared to be present in canalicular membranes from rat liver [12], in accordance with the concentrative guanosine transport activity found in this study. This apparent discrepancy would be explained if this carrier is present at the canalicular plasma membrane but below the detection level of the antibody or if another concentrative nucleoside transporter, for instance the recently cloned CNT3 isoform [7], is expressed. Although this study does not rule out the possibility of direct targeting of CNT1 into the apical membrane of the hepatocyte, it FIG. 6. Immunohistochemistry of rCNT1 and caveolin in hepatocytes. Immunofluorescence was used for the histochemical analysis of CNT1 (A) and caveolin (B) in paraformaldehyde-fixed sections of rat liver. Sections were incubated with respective primary antibodies, followed by anti-rabbit Oregon-green labeling and Texas-redsecondary antibodies for CNT1 and caveolin, respectively. A and B show the same section stained with the two antibodies. A similar pattern of CNT1 and caveolin distribution is found at the basolateral membrane level (white arrows). n, nucleus.

location, CNT1 has been identified in transcytotic endocytic fractions and in the early/sorting endosomes. RRC contains vesicles of different sizes and tubular membranes and has recycling receptors and transcytotic and caveolar markers [36]. These endosomal fractions from rat liver have been characterized morphologically and biochemically [17, 18, 36 –39], as also shown in this study with a variety of markers. The transport kinetics through these isolated fractions of several endocytosed ligands such as LDL, asialofetuin, pIgA, transferrin, or EGF has also been studied in detail [20, 40]. More recently, a new plasma membrane fraction derived from the sinusoidal plasma membrane but enriched in caveolin (CEF) has been characterized [18, 25]. CEF contained most of the caveolae markers described [25] and, as shown here, it is highly enriched in CNT1 transporters.

FIG. 7. Model of rCNT1 pathways in the hepatocyte. This model is consistent with the distribution pattern of rCNT1 in endosomes and plasma membrane domains and suggests that rCNT1 is basically located close to caveolae, in contrast to rCNT2, which appears to be located in other microdomains of the sinusoidal membrane, in which it is fully active. rCNT1 is inserted on the canalicular side of the hepatocyte following the indirect transcytotic pathway.

84

DUFLOT ET AL.

shows that CNT1 follows, at least, the indirect route for insertion. These results open a new possibility for the regulation of nucleoside transport activity in liver parenchymal cells, which may rely upon internalization, transcytosis, and insertion of transporter proteins at different domains of the plasma membrane. Interestingly, the same vesicular structures that transport CNT1 also contain molecules involved in the MAPK signal transduction pathway [36]. Indeed, portal injection of EGF leads to rapid translocation of Raf-1 from CEF into different endocytic compartments (RRC and CURL), thus resulting in Mek phosphorylation and, eventually, MAPK activation. The possibility that growth factors modulate nucleoside transport in hepatocytes is now being studied. In conclusion, CNT1 and CNT2 subcellular localization and trafficking in hepatocytes appear to be differentially regulated. CNT1 follows the transcytotic pathway to the canalicular membrane. This is the first evidence of transcytosis for a plasma membrane transporter in liver parenchymal cells. This work was supported by Grants SAF99-0115, SAF2002-717, and 2FD97-1268 to MPA and PM99-0166 to CE. We thank the Dr. Marta Taule´ s for her technical assistance in immunoelectron microscopy, Dr. Marta Camps for her helpful discussions, and Robin Rycroft for his editorial help.

REFERENCES 1.

Griffith, D. A., and Jarvis, S. M. (1996). Nucleoside and nucleobase transport systems of mammalian cells. Biochim. Biophys. Biochim. Biophys Acta 1286, 153–181.

2.

Wang, J., Schaner, M. E., Thomassen, S., Su, S. F., PiquetteMiller, M., and Giacomini, K. M. (1997). Functional and molecular characteristics of Na(⫹)-dependent nucleoside transporters. Pharm. Res. 14, 1524 –1532.

3.

4.

Pastor-Anglada, M., Felipe, A., and Casado, F. J. (1998). Transport and mode of action of nucleoside derivatives used in chemical and antiviral therapies. Trends Pharmacol. Sci. 19, 424 – 430. Baldwin, S. A., Mackay, J. R., Cass, C. E. and Young, J. D. (1999). Nucleoside transporters: Molecular biology and implications for therapeutic development Mol. Med. Today 5, 216 –224.

5.

Felipe, A., Valde´ s, R., del Santo, B., Lloberas, J., Casado, J. and Pastor-Anglada, M. (1998). Na⫹-dependent nucleoside transport in liver: Two different isoforms from the same gene family are expressed in liver cells. Biochem. J. 330, 997–1001.

6.

del Santo, B., Valde´ s, R., Mata, J., Felipe, A., Casado, F. J. and Pastor-Anglada, M. (1998). Differential expression and regulation of nucleoside transport systems in rat liver parenchymal and hepatoma cells. Hepatology 28, 1504 –1511.

7.

Ritzel, M. W. L., Ng, A. M. L., Yao, S. Y. M., Graham, K., Loewen, S. K., Smith, K. M., Ritzel, R. G., Mowles, D. A., Carpenter, A., Chen, X. Z., Karpinski, E., Hyde, R. J., Baldwin, S. A., Cass, C. E., and Young, J. D. (2001). Recent molecular advances in studies of the concentrative Na⫹-dependent nucleoside transporter (CNT) family: Identification and characterization of novel human and mouse proteins (hCNT3 and mCNT3)

broadly selective for purine and pyrimidine nucleosides (system cib). J. Biol. Chem. 276, 2914 –2927. 8.

Evans, W. H. (1980). A biochemical dissection of the functional polarity of the plasma membrane of the hepatocyte. Biochim. Biophys. Acta 604, 27– 64.

9.

Bartles, J. R., Feracci, H. M., Stieger, B., and Hubbard, A. L. (1987). Biogenesis of the rat hepatocyte plasma membrane in vivo: Comparison of the pathways taken by apical and basolateral proteins using subcellular fractionation. J. Cell Biol. 105, 1241–1251.

10.

Schell, M. J., Maurice, M., Stieger, B., and Hubbard, A. L. (1992). 5⬘ nucleotidase is sorted to the apical domain of hepatocytes via an indirect route. J. Cell Biol. 119, 1173–1182.

11.

Kipp, H., and Arias, I. M. (2000). Newly synthesized canalicular ABC transporters are directly targeted from the Golgi to the hepatocyte apical domain in rat liver. J. Biol. Chem. 275, 15917–15925.

12.

Che, M., Nishida, T., Gatmaitzan, Z., and Arias, I. M. (1995). Primary structure and functional expression of a cDNA encoding the bile canalicular, purine-specific Na(⫹)-nucleoside cotransporter. J. Biol. Chem. 270, 13596 –13599.

13.

Hamilton, S. R., Yao, S. Y. M., Ingram, J. C., Hadden, D. A., Ritzel, M. W. L., Gallagher, M. P., Henderson, P. F., Cass, C. E., Young, J. D. and Baldwin, S. A. (2001). Subcellular distribution and membrane topology of the mammalian concentrative Na⫹nucleoside cotransporter rCNT1. J. Biol. Chem. 276, 27981– 2798.

14.

Ohno, H., Tomemori, T., Nakatsu, F., Okasaki, Y., Aguilar, R. C., Foelsch, H., mellman, I., Saito, T., Shirasawa, T., and Bonifacino, J. S. (1999). Mu 1B, a novel adaptor medium chain expressed in polarized epithelial cells. FEBS Lett. 449, (2–3), 215–220.

15.

Koivisto, U-M., Hubbart, A. L., and Mellman, I. (2001). A novel cellular phenotype for familial hypercholesteriolemia due to a defect in polarized targetting of LDL receptor. Cell 105, 575– 585.

16.

Pastor-Anglada, M., Remesar, X., and Bourdel, G. (1987). Alanine uptake by liver at midpregnancy in rats. Am. J. Physiol. 252, E408 –E413.

17.

Neville, D. M., Jr. (1960). The isolation of a cell membrane fraction from a rat liver. J. Biophys. Biochem. Cytol. 8, 413– 422.

18.

Belcher, J. D., Hamilton, R. L., Brady, S. E., Ja¨ ckle, S., Schneider, W. J. and Havel, R. J. (1987). Isolation and characterization of three endosomal fractions from the liver of estradioltreated rats. Proc. Natl. Acad. Sci. USA 84, 6785– 6789.

19.

Calvo, M., and Enrich, C. (2000). Biochemical analysis of a caveolae-enriched plasma membrane fraction from rat liver. Electrophoresis 21, 3386 –3395.

20.

Ja¨ ckle, S., Runquist, E. A., Miranda-Brady, S., and Havel, R. J. (1991). Trafficking of the epidermal growth factor receptor and transferrin in three hepatocytic endosomal fractions. J. Biol. Chem. 266, 1396 –1402.

21.

Ruiz-Montasell, B., Casado, F. J., Felipe, A., and PastorAnglada, M. (1992). Uridine transport in basolateral plasma membrane vesicles from rat liver. J. Membrane Biol. 128, 227– 233.

22.

Dragan, Y., Valde´ s, R., Go´ mez-Angelats, M., Felipe, A., Casado, F. J., Pitot, H. C., and Pastor-Anglada, M. (2000). Selective loss of nucleoside carrier expression in rat hepatocarcinomas. Hepatology 32, 239 –246.

23.

Pol, A., Ortega, D., and Enrich, C. (1997). Identification and distribution of proteins in isolated endosomal fractions of rat

CNT1 FOLLOWS THE HEPATIC TRANSCYTOTIC PATHWAY

24.

25.

26.

27.

28.

29.

30.

31.

liver: Involvement in endocytosis, recycling and transcytosis. Biochem. J. 323, 435– 443. Stieger, B., Hagenbuch, B., Landmann, L., Hochli, M., Schroeder, A., and Meier, P. J. (1994). In situ localization of the hepatocytic Na⫹/taurocholate cotransporting polypeptide in rat liver. Gastroenterology 107, 1781–1787. Pol, A., Calvo, M., Lu, A., and Enrich, C. (1999). The “earlysorting” endocytic compartment of rat hepatocytes is involved in the intracellular pathway of caveolin-1 (VIP-21). Hepatology 29, 1848 –1857. Babbit J., Trigatti, B., Rigotti, A., Smart, E. J., Anderson, R. G. W., Xu, S., and Kriger, M. (1997). Murine SR-BI, a high density lipoprotein receptor that mediates selective lipid uptake, is N-glycosylated and fatty acid acylated and colocalizes with plasma membrane caveolae. J. Biol. Chem. 272, 13242– 13249. Graf, G. A., Connell, P. M., van der Westhuyzen, D. R. and Smart, E. J. (1999). The class B, type I scavenger receptor promotes the selective uptake of high density lipoprotein cholesterol ethers into caveolae. J. Biol. Chem. 274, 12043–12048. Schnitzer, J. E., McIntosh, D. P., Dvorak, A. M., Liu, J., and Oh, P. (1995). Separation of caveolae from associated microdomains of GPI-anchored proteins. Science 269, 1435–1439. Luton, F., Verge´ s, M., Vaerman, J.-P., Sudol, M., and Mostov, K. E. (1999). The SRC family of protein tyrosine kinase p62 yes controlsymeric IgA transcytosis in vivo. Mol. Cell 4, 627– 632. Verge´ s, M., Havel, R. J., and Mostov, K. E. (1999). A tubular endosomal fraction from rat liver: Biochemical evidence of receptor sorting by default. Proc. Natl. Acad. Sci. USA 96, 10146 – 10151. Pol, A., Calvo, M., Lu, A., and Enrich, C. (2000). EGF triggers caveolin redistribution from the plasma membrane to the early/ sorting endocytic compartment of hepatocytes. Cell Signal. 12, 537–540.

Received February 13, 2002 Revised version received June 24, 2002 Published online October 11, 2002

85

32.

Calvo, M., Tebar, F., Lopez-Iglesias, C., and Enrich, C. (2001). Morphological and functional characterization of caveolae from liver hepatocytes. Hepatology 33, 1259 –1269.

33.

Mercader, J., Go´ mez-Angelats, M., del Santo, B., Casado, J., Felipe, A., and Pastor-Anglada, M. (1996). Nucleoside uptake in rat liver parenchymal cells. Biochem. J. 317, 835– 842.

34.

Ruiz-Montasell, B., Martinez-Mas, J. V., Enrich, C., Casado, F. J., Felipe, A., and Pastor-Anglada, M. (1993). Early induction of Na(⫹)-dependent uridine uptake in the regenerating rat liver. FEBS Lett. 316, 85– 88.

35.

del Santo, B., Tarafa, G., Felipe, A., Casado, F. J. and PastorAnglada, M. (2001). Developmental regulation of the concentrative nucleoside transporters CNT1 and CNT2 in rat liver. J. Hepatol. 34, 873– 880.

36.

Pol, A., Calvo, M.,and Enrich, C. (1998). Isolated endosomes from quiescent rat liver contain the signal transduction machinery. Differential distribution of activated Raf-1 and Mek in the endocytic compartment. FEBS Lett. 441, 34 –38.

37.

Enrich, C., Pol, A., Calvo, M., Pons, M.,and Ja¨ ckle, S. (1999). Dissection of the multifunctional “receptor-recycling” endocytic compartment of hepatocytes. Hepatology 30, 1115–1120.

38.

Calvo, M., Pol, A., Lu, A., Ortega, D., Pons, M., Blasi, J., and Enrich, C. (2000). Cellubrevin is present in the basolateral endocytic compartment of hepatocytes and follows the transcytotic pathway after IgA internalization. J. Biol. Chem. 275, 7910 –7917.

39.

Anderson, R. G. W. (1998). The caveolae membrane system. Annu. Rev. Biochem. 67, 199 –225.

40.

Enrich, C., Ja¨ ckle, S., and Havel, R. J. (1996). The polymeric immunoglobulin receptor is the major calmodulin-binding protein in an endosome fraction from rat liver enriched in recycling receptors. Hepatology 24, 226 –232.